1. Field of the Invention
The present invention relates to a laminated semiconductor substrate and an optical semiconductor element.
2. Related Background Art
A long wavelength (from 1.25 μm to 1.6 μm) optical semiconductor element, especially 1.3 μm wavelength band (from 1.25 μm to 1.35 μm) optical semiconductor element attracts attention as a light emitting element or a photodetector element for optical communication. In the prior art, the semiconductor light emitting element of this long wavelength has a structure in which an active layer (light emitting layer) of an InGaAsP compound semiconductor is formed on an InP substrate (supporting substrate). In this structure, lattice mismatch between the active layer and the substrate is small, and the element is easy to manufacture.
The semiconductor light emitting element using the InP substrate is, however, expensive because the InP substrate is expensive. The element using the InP substrate has a bad temperature characteristic. Therefore, the problem with the semiconductor light emitting element using the InP substrate is that it is expensive and has a bad temperature characteristic. In order to provide inexpensive optical semiconductor elements excellent in temperature characteristic, the development of elements that comprise a GaAs substrate and an active layer of an InGaAs compound semiconductor has been promoted. The GaAs substrate is generally used in an optical semiconductor element for a wavelength of 0.98 μm, and has the advantages of being inexpensive and easy to work. In order to provide inexpensive optical semiconductor elements for a long wavelength, using those advantages, the development of elements using GaAs substrates has been promoted.
However, the problem with the semiconductor light emitting element using the GaAs substrate for a long wavelength also is that its light emission intensity is low. This is because the lattice constant of the InGaAs active layer is large compared to the GaAs substrate and the lattice mismatch between them is large. Therefore, cracks are liable to occur in the active layer.
In somewhat more detail, as the In composition of InGaAs increases, the bandgap wavelength increases. Therefore, in order to provide an optical semiconductor element using an InGaAs active layer for a long wavelength, the In composition of the active layer should increase. In addition, in order to provide an optical semiconductor element for a long wavelength, the thickness of the active layer should be large because if the thickness of the active layer decreases, the wavelength of the element would be short due to a quantum effect. Therefore, in order to provide an optical semiconductor element for a long wavelength, using the InGaAs active layer, an active layer having a high In composition and a large thickness should be formed. However, when the In composition of the active layer increases, the lattice mismatch between the InGaAs active layer and the GaAs substrate would increase. This causes to decrease the critical thickness of the active layer for the substrate based upon equilibrium theories. If the active layer has a larger thickness than the critical thickness, a large number of cracks would usually occur in the active layer. As a result, in the conventional optical semiconductor element using the GaAs substrate for a long wavelength, a large number of cracks would occur in the active layer to thereby extremely decrease the light emission intensity.
In order to prevent cracks from occurring in the active layer, a method of mixing nitrogen (N) into the active layer to thereby decrease the lattice constant of the active layer was used. This method increases the critical thickness of the active layer for the substrate based upon equilibrium theories. In this method, however, the concentration of nitrogen in the active layer increases to thereby decrease the light emission intensity. Consequently, this method cannot provide a sufficiently high light emission intensity. As another method of preventing the occurrence of cracks in the active layer, it has been tried to provide a buffer layer between the substrate and the active layer. This method intends to prevent the number of cracks from increasing in the active layer even when the active layer has a larger thickness than its critical thickness for the substrate based upon the equilibrium theories. In this method, a certain effect is produced when the active layer and the substrate are composed of materials the lattice mismatch between which is small. However, in the conventional buffer layer, the number of cracks occurring in the active layer cannot be sufficiently reduced when there is a large lattice mismatch between the substrate and the active layer. Thus, even by using such buffer layer, the light emission intensity cannot sufficiently increase.
As described above, the problem with the conventional semiconductor light emission element using the GaAs substrate for a long wavelength is that the light emission intensity is low.
As with the optical semiconductor element using the GaAs substrate for a long wavelength, the problem with the laminated semiconductor substrate is that as the lattice mismatch between the substrate and the semiconductor layer increases, high properties cannot be obtained.
More particularly, generally, a semiconductor element is produced by using a laminated semiconductor substrate that includes a semiconductor substrate hundreds of μm thick and a semiconductor layer several μm thick formed on the substrate. When the lattice mismatch between the substrate and the semiconductor layer increases in this laminated semiconductor substrate, the number of cracks occurring in the semiconductor layer is liable to increase. This is because as the number of the lattice mismatch increases, the critical thickness of the semiconductor layer for the substrate based upon the equilibrium theories decreases whereas when the semiconductor layer is formed so as to have a thickness of more than the critical thickness, cracks are liable to occur. Therefore, in the semiconductor element, the lattice mismatch between the substrate and the semiconductor layer should be reduced as much as possible.
The substrates generally used in the semiconductor elements are limited to substrates made of Si, GaAs, InP, GaP and InAs, respectively, and all substrates having their respective proper lattice constants cannot be used. Therefore, when the formation of a semiconductor layer having a specified function on a substrate is tried, there occurs a lattice mismatch between the semiconductor layer and the substrate in many cases. In order to reduce the number of cracks occurring in the semiconductor layer even when there is such a lattice mismatch, a method of growing a buffer layer having a lattice constant between those of the substrate and the semiconductor layer has been tried. This method is disclosed, for example, in Published Japanese Patent Application Hei 7-94524. In this method, when the lattice mismatch between the substrate and the semiconductor layer is small the buffer layer absorbs a considerable number of cracks in the semiconductor layer to thereby reduce the number of cracks therein.
In this method, however, when the lattice mismatch between the substrate and the semiconductor layer increases, the buffer layer itself would be subjected to distortion to thereby store distortion energy for elastic deformation in the entire buffer layer. When the distortion energy becomes larger than the crack energy, it is difficult to prevent the occurrence of cracks. Therefore, when the lattice mismatch between the substrate and the semiconductor layer increases in the conventional buffer layer, it is difficult to alleviate the influence of the lattice mismatch. More specifically, when the lattice mismatch is not less than 1%, it is difficult to alleviate the influence of the lattice mismatch and the number of cracks in the semiconductor layer cannot be reduced sufficiently.
As described above, the problem with the laminated semiconductor substrate is that when the lattice mismatch between the substrate and the semiconductor layer increases, a high property cannot be obtained. This is because it is difficult to increase the thickness of the semiconductor layer beyond its critical thickness for the substrate based upon the equilibrium theories and reduce the number of cracks occurring thereon.